Research map / Label-efficient and robust learning / Recognition and long-tail
Semantic Proto-Role Labeling
Adam Teichert, Adam Poliak, Benjamin Van Durme, Matthew Gormley
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add pairwise property (second-order) factors (ablation) | SPRL · gold syntax | F1: 80.9 -> 81.7 | Improved |
| observe gold SRL ArgN roles (ablation) | SPRL · gold syntax | F1: 80.9 -> 81.5 | Improved |
| observe gold SRL roles / joint SRL+SPRL (ablation) | SPRL · gold syntax | F1: 81.7 -> 81.8 | No clear effect |
| remove syntax features (ablation) | SPRL · SPRL F1 | F1: 81.7 -> 80.8 | No clear effect |
| remove syntax features (ablation) | ArgN SRL · PropSmall | accuracy: 87.0 -> 82.7 | Hurt |
| remove syntax features (ablation) | SFT SRL · PropSmall | accuracy: 79.1 -> 79.1 | No clear effect |
| add observed gold SPRL features (ablation) | ArgN SRL · PropSmall, gold syntax | accuracy: 87.0 -> 87.7 | Improved |
| add observed gold SPRL features (ablation) | SFT SRL · PropSmall, gold syntax | accuracy: 79.1 -> 80.2 | Improved |
| joint modeling with SPRL (train-time only) (ablation) | ArgN SRL · PropSmall, no syntax | accuracy: 82.7 -> 84.4 | Improved |
| joint modeling with SPRL (train-time only) (ablation) | ArgN SRL · PropSmall, gold syntax | accuracy: 87.0 -> 86.5 | Hurt |
| add second-order SPRL features (ablation) | SRL|SPRL · PropSmall, gold syntax ArgN | accuracy: 87.7 -> 86.8 | Hurt |
| use SFT labels instead of ArgN (ablation) | SRL · PropSmall, gold syntax | accuracy: 87.0 -> 79.1 | Hurt |
Closest papers by meaning
- Dependency or Span, End-to-End Uniform Semantic Role Labeling (AAAI 2019 · similarity 0.86)
- Span-Based Semantic Role Labeling with Argument Pruning and Second-Order Inference (AAAI 2022 · similarity 0.86)
- Dependency Tree Representations of Predicate-Argument Structures (AAAI 2016 · similarity 0.85)
- Encoder-Decoder Based Unified Semantic Role Labeling with Label-Aware Syntax (AAAI 2021 · similarity 0.85)
- Frame Semantic Role Labeling Using Arbitrary-Order Conditional Random Fields (AAAI 2024 · similarity 0.84)
- The Illusion of Role Separation: Hidden Shortcuts in LLM Role Learning (and How to Fix Them) (ICML 2025 · similarity 0.83)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
